GCN4 ia a transcriptional activator of amino acid biosynthetic genes in S. cerevisiae that is regulated translationally by short open reading frame (uORFs) in the GCN4 mRNA leader. Ribosomes must translate uORF1, resume scanning and bypass the start codons at uORFs 2-4 in order to translate GCN4. This occurs during amino acid starvation because reinitiation of translation is inefficient under these conditions. The GCD1, GCD2, GCD6, and GCD7 proteins are required for translational repression of GCN4 and have general functions in translation initiation. They are components of a complex that stably interacts with a portion of the initiation factor 2 (eIF-2) in yeast cells. This complex is the yeast equivalent of initiation factor eIF-2B which catalyzes GDP-GTP exchange on eIF-2. eIF-2B activity is downregulated in mammalian cells by phosphorylation of the a subunit of eIF-2. We have demonstrated that the protein kinase GCN2 stimulates translation of GCN4 in starved east cells by phosphorylation of eIF-2a. This is expected to inhibit eIF-2B function and reduce the levels of eIF-2 activity in the cell, allowing ribosomes scanning downstream from uORF1 to skip over uORFs 2-4 and reinitiate at GCN4 instead. Two mammalian eIF-2alpha kinases, DAI and HCR, substitute for GCN2 and stimulate GCN4 translation in yeast cells. Thus, the regulation of GCN4 expression is a gene-specific example of translational control by phosphorylation of eIF-2. We have found that purine starvation activates GCN2 kinase function and that this response is required for wild-type levels of purine biosynthesis under starvation conditions, indicating that the GCN4 regulatory mechanism is more global than was previously imagined. Ribosomal protein (rp) genes are transcriptionally repressed under amino-acid starvation conditions and this repression requires the binding site for the regulatory protein RAP1 in the rp promoter. Mutations in RAP1 and in several chromosomal genes have been identified that alter the regulation of rp gene expression in response to growth rate and amino acid levels. These results provide in vivo evidence for the involvement of RAP1 in controlling rp gene transcription and identify new components of this signal transduction mechanism.